Between Reasonable Doubt and Reasonable Range

The term ‘Reasonable Doubt’ is a key concept in the legal system. Judging people is a typical decision making process under uncertainty, with drastic consequences to the individual and also to the whole society.  The essence of the idea is to minimize the cases where innocent people are punished, while also limiting the cases where guilty people get away.

For me this is one example where the best solution to a conflict is a compromise! I don’t have another solution to the inherent conflict in the quest to reduce crime while doing justice to every human being.

The basic assumption we would like to challenge, but unable to, is “never say I know”. Goldratt coined it (was he the first???) and it certainly applies to judges and juries who have to pass a judgment.

In management of organizations the drama is usually on a lower level. Very few decisions could cause a comparable damage of jailing an innocent person, or even setting a murderer free.

Yet, every day every CEO has to decide upon matters that might cause considerable damage, but might also cause considerable benefits. Like in the justice system one might claim that the solution is a compromise, but for managing an organization we are able, to my mind, to resolve of the conflict of take a risky action versus do not take the risky action.  Because the organization can tolerate decisions leading to certain damages as long as the accumulation of all decisions is highly positive.  This is not the case with the justice system.

In previous posts I have already mentioned the concept of “the reasonable range”, meaning considering the result of a “reasonable pessimistic forecast” versus “reasonable optimistic forecast”, uncovering the damage/benefit of each and making the final decision knowing that the real impact would fall, in most cases, somewhere between the two.

There is one insight coming from the legal system that is especially important:

The damage of a certain potential result is so big that we have to prevent it at any cost!

The concept of “above any reasonable doubt” means that no matter what horrible deed the defendant might have done – we cannot allow the damage of punishing an innocent person.

The generic insight is that any decision that might, even only in relatively infrequent cases, cause big damage then that decision should be rejected. If there is a chance, say of 5%, that the defendant is not guilty, meaning 95% that he/she is, then the right decision is to acquit.  The injustice done to an individual is perceived much more than 10 times the damage to society. Even 1% of doubt is usually considered reasonable.

Translating the insight to management environments:

Managers have to draw a red-line of potential damage resulting from their decisions and reject any decision that might penetrate the red-line, no matter what are the benefits.

People tend to make a distinction between losing a dollar and making a dollar, preferring not to lose. With small numbers this connotation is flawed and pushes us to wrong decisions.  However, I’d give up making $100M if the decision might cause me $10M loss, because the damage to my life from losing $10M is far greater than the benefits of gaining $100M.  For a billionaire this is the wrong decision, because for the billionaire the damage of losing $10M is much lower than for me.

So, unlike the justice system that does not consider the balance of how many free guilty people are out of jail, in management the financial status of the organization is an important parameter in assessing the potential damage of a decision.

There is another interesting insight coming from the legal concept of a reasonable doubt:

The judgment whether there is a reasonable doubt is based on intuition!

In other words, there is no strict method to deduce whether the evidence still leaves a reasonable doubt. Even though the justice system is very strict with the process of justice, the word ‘reasonable’ leaves the whole burden of the decision to convict or acquit on the shoulders of a human being using his intuition.

In management there is tendency to adore the mathematics of optimization, even though not to the same level of the academy, for which optimization is the key.  I think we should learn from the legal system that eventually you need your intuition to cover for lack of good information.

But, my observation is that both the legal and the managerial systems lack the ability to support the basic intuition about doubt and risk. There are enough biases that we, human beings, have when we intuitively assess uncertainty/risk/doubts.  Prof. Kahneman points to several of those biases in his book Thinking Fast and Slow.  Understanding better the key concepts and principles of Probability Theory could help us analyze (rather them deciding by an instinct) the conflict, still using our intuition, but also controlling it by using the principles.

I recently read an article by a lawyer who is also a professor for law in Israel that argued that every single piece of evidence has to be judged on its own terms and when there is a reasonable doubt about that particular piece it should be ignored. In other words, according to that professor there is no additional weight to the overall evidence in judging whether there is a reasonable doubt.  I sharply disagree with such an assertion that clashes with the basic logic of Mathematics.  I try to find the bridge between logic and reality where we lack the full information required by the Mathematical and Statistical models.

That said, the Mathematical logic defines the concept of “dependence” and “independence” between variables, like between different pieces of evidence or the demand in different locations. My observation of reality is that in too many cases there are “partial dependencies” between the variables, which means the effect of reduced risk due to aggregation is still there, but smaller than what we sometimes are led to believe. Once we understand the basic logic we can improve the impact of our intuition and this would yield very high value.